MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Transferring Beam Navigation Behavior From Human to Robot: An Evidence Driven Decision Making Model for Liver SBRT

Y Sheng*, W Wang, R Li, C Wang, J Zhang, X Li, H Stephens, Q Wu, F Yin, Y Ge, Q Wu, Duke University Medical Center, Durham, NC

Presentations

(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: To develop decision making model to learn human planner’s beam navigation behavior for beam angle/arc angle selection for liver SBRT.


Methods: A total of 27 liver SBRT/HIGRT patients (10 IMRT, 17 VMAT/DCA) were included in this study. A dosimetric budget index was defined for each beam angle/control point considering the body as well the liver tissue. Optimal beam angle setting (beam angles for IMRT and start/stop angle for VMAT/DCA) was determined by minimizing the loss function defined as the sum of total dosimetric budget index and beam span penalty function. Leave-one-out validation was exercised on all 27 case while hyperparameters in the loss function was tuned in nested cross validation. To compare the efficacy of the model, an evidence guided plan (EG-plan) was generated using automatically generated beam setting together with original optimization constraints in the clinical plan. EG-plan was normalized to the same PTV V100% as clinical plan. Dosimetric endpoints including PTV D98%, D2%, liver V20Gy and total MU were compared between two plan groups. Wilcoxon Signed-Rank test was performed with the null hypothesis that no difference exists between two groups.


Results: Beam setting prediction is instantaneous. Mean PTV D98% was 91.3% and 91.3% (p=0.164), while mean PTV D2% was 107.9% and 108.1% (p=0.566) for clinical plan and EG-plan respectively. Liver V20Gy showed no significant difference (p=0.590) with 23.3% for clinical plan and 23.4% for EG-plan. Total MU is comparable (0.256) between clinical plan (2389.6) and EG-plan (2319.6).


Conclusion: The evidence driven beam setting model yielded similar plan quality as hand-crafted clinical plan. It is capable of capturing human’s reasoning in beam selection decision making. This model could facilitate decision making for beam angle selection choices while eliminating lengthy trial-and-error process of adjusting beam setting during liver SBRT treatment planning.

Funding Support, Disclosures, and Conflict of Interest: This study is partially supported by NIH grant R01CA201212 and a master research grant by Varian Medical Systems

Keywords

Treatment Planning

Taxonomy

TH- External Beam- Photons: Treatment planning using machine learning/Knowledge Based Planning/automation

Contact Email